[Cluster-devel] [PATCH] dlm: bz#251179 Fix lowcomms socket closing

[Steven Whitehouse <swhiteho@redhat.com>]

Hi,

Now in the -nmw git tree. Thanks,

Steve.

On Mon, 2007-08-20 at 15:13 +0100, Patrick Caulfield wrote:
> This patch fixes the slight mess made in lowcomms closing by previous patches
> and fixes all sorts of DLM hangs.
> 
> Signed-Off-By: Patrick Caulfield <pcaulfie@redhat.com>
> 
> plain text document attachment (lowcomms-close-patch.stevew)
> --- gfs2-2.6-nmw/fs/dlm/lowcomms.c.orig	2007-08-20 15:05:48.000000000 +0100
> +++ gfs2-2.6-nmw/fs/dlm/lowcomms.c	2007-08-20 15:07:00.000000000 +0100
> @@ -334,18 +334,8 @@ static void close_connection(struct conn
>  		con->rx_page = NULL;
>  	}
>  
> -	/* If we are an 'othercon' then NULL the pointer to us
> -	   from the parent and tidy ourself up */
> -	if (test_bit(CF_IS_OTHERCON, &con->flags)) {
> -		struct connection *parent = __nodeid2con(con->nodeid, 0);
> -		parent->othercon = NULL;
> -		kmem_cache_free(con_cache, con);
> -	}
> -	else {
> -		/* Parent connections get reused */
> -		con->retries = 0;
> -		mutex_unlock(&con->sock_mutex);
> -	}
> +	con->retries = 0;
> +	mutex_unlock(&con->sock_mutex);
>  }
>  
>  /* We only send shutdown messages to nodes that are not part of the cluster */
> @@ -731,6 +721,8 @@ static int tcp_accept_from_sock(struct c
>  			INIT_WORK(&othercon->swork, process_send_sockets);
>  			INIT_WORK(&othercon->rwork, process_recv_sockets);
>  			set_bit(CF_IS_OTHERCON, &othercon->flags);
> +		}
> +		if (!othercon->sock) {
>  			newcon->othercon = othercon;
>  			othercon->sock = newsock;
>  			newsock->sk->sk_user_data = othercon;